The first phase of this project aims to deliver an evolving, experimental low-cost, rapid deployment customer service chatbot for a local astronomical society, featuring Grounding LLM (e.g. RAG), embedding, Agentic AI, function calling to external service.

• Low cost
• Rapid deployment

• Azure AI Search (Standard S0 tier)
• Azure Cosmos DB for NoSQL (Free Tier)
• Azure Apps (F1 Tier)
• Azure OpenAI models
  • gpt-4.1 model for chatbot
  • text-embedding-3-small for embedding
• Azure Data Factory for regular data download and indexing trigger via REST API

https://aiforastro.azurewebsites.net/

• Grounding LLM (via RAG)
• Embedding on vectorization

• Agentic AI
• Function calling to third party service